Predicting Voting Patterns in India From Social Media Political Campaigns

Nearly five years have passed since I started my study. Now, elections are around the corner again for most of the states that I analysed. Will I put the Singh-Srivastava Vote Predictor, a statistical model that correlates political communication on Twitter with actual vote share, through its paces once more? It would certainly be the right thing to do. There is nothing more robust than testing and retesting a model at regular intervals to measure its efficacy. But this is a humongous exercise, so I will not be doing it again. 

The tech policy of the platform has also changed, with Musk relying more on automation and less on human interface to manage Twitter. Several alterations have also been made to the platform’s APIs. This affects the significant community of developers who now must rewire and tweak their processes of data acquisition and curation.

My study was completed long before these changes on the Twitter platform were even a gleam in Elon Musk’s eye. Nevertheless, even after these changes, the Singh-Srivastava Vote Predictor stands firm because it simply measures the engagement that politicians and political parties receive on Twitter and correlates this online engagement with a predictive vote share offline. 

The implications of this are huge because it shows that the impact of an online campaign can genuinely be measured in real life. However, this does not mean that online campaigns can be a substitute for offline campaigns. For example, in the months preceding the 2019 Lok Sabha elections, Rahul Gandhi received better average engagement on Twitter than Narendra Modi. But that didn’t translate into a win for the Congress party because their offline campaign strategy was weak compared to that of the BJP. Online campaigns can add value to the brand of a political actor, but they need to be complemented with similar offline strategies, as well as the mass mobilisation of the party cadre and supporters. 

My finding also shows that political actors cannot use Twitter randomly, or only in the run-up to elections. If they don’t engage with their audience 24/7, they lose their traction. Keep in mind that what matters is engagement, not the number of tweets. As I showed in Chapter 3, more tweets do not necessarily mean more engagement.

The other significant implication of my study is that the digital divide between richer and poorer states in India is closing. The seven states I analysed ranged from the more developed Maharashtra and Haryana to the poorer Chhattisgarh and Jharkhand; yet by measuring the engagement of the state politicians and political parties in their Twitter campaigns, I was successfully able to predict party vote share even in less developed states. The primary reason for this is the digital revolution of 2016 that resulted in a truly digital consumption society. The number of internet users in India went up exponentially between 2014 (approximately 25 crore) to 2019 (64 crore), which was more than half of the 91.1 crore people eligible to vote in the 2019 Lok Sabha elections. 

This was why the findings of my study so accurately reflected the results achieved by traditional offline campaigns. It is clear now that political actors cannot afford to ignore Twitter as a tool for their activities. The blurring of the online and offline worlds will probably continue to increase, making social media a true part of the real world and changing the nature of politics even more.

This means there is no escaping social media for political leaders. The more people identify with them online, the higher their chances of engagement and the potential conversion of this engagement into vote shares during elections. Engagement means involvement in constant two-way communication with users, exchanging ideas and information, and thus helping grow a support base far and wide. Consistency of thought is also key to ensuring that political leaders can command a loyal online support base. Political leaders need to be consistent with their views and make sure that they are not limiting their followers to just their or their party views but exposing them to a global space where they can gain knowledge and expand their social network to include others who follow the same leaders or leaders with similar functioning traits across the world. Consistent effort on the part of political leaders to constantly lend an ear to their support base and keep them informed of the latest practices/information, etc. can be used to their political advantage at the right time during elections. They also need to engage with followers who react negatively to their posts. This is important because it shows that they are open to criticism and willing to discuss issues, and it also puts them in a position to convince angry followers of their intentions. 

Likewise, political parties need to up their game and actively engage users 24/7. Blanking out between elections will no longer work. Political leaders need to assign professionals to run their party Twitter handles via which the audience can be engaged in a positive manner throughout the year.

Excerpted with permission from The Online Effect: Deciding X To Predict Election Outcomes (Bloomsbury) by Sanjeev Singh. 

Disclaimer: The author and publisher of the book are solely responsible for the contents of the book or any excerpt derived therefrom. NDTV shall not be responsible or liable for any claims arising from the contents of the book including any claims of defamation, infringement of intellectual property rights or any other right of any third party or of law.